Non-intrusive hourly occupancy detection in residential buildings using remotely readable water meter data: Validation and large-scale analysis

Published
Data Science
Energy
Authors

M. Schaffer

J.E. Vera-Valdés

A. Marszal-Pomianowska

Published

2025

Abstract

This study presents a novel, unsupervised algorithm to detect hourly and daily occupancy patterns in residential buildings using data from remotely readable smart water meters. Validated with ground-truth data from Danish households, the method achieves high accuracy for daily presence and reliably detects absences longer than three hours. Unlike intrusive or costly sensor-based approaches, this water meter-based method is privacy-friendly, scalable, and robust.

The algorithm outperforms leading electricity meter-based solutions and can be applied on a large scale—demonstrated with a dataset from over 2,600 homes. The findings have significant implications for building energy modeling, urban energy policy, and smart grid integration, providing practical and accessible occupancy estimates for research and demand response strategies.

Download

The paper can be downloaded here.